Re: Where can I find information on the computers used by the Apollo missions?

Blaine,

When I received your question about the getting information on the computers used in the Apollo program I immediately thought of a friend of mine, Dr. Warren F. Davis, who did programming development work for the Apollo program. (As I recall, he did a lot (all?) of the work on the program that would separate the Command Module from the Saturn booster in case something went terribly wrong in the first few seconds of launch and on the program that brought the Command Module back to earth from orbit at the end of the mission.) I wrote to him, forwarding your question. Here's what he said:

Here's what I can offer toward an answer to the question about the Apollo computers. Sorry that it's a bit long-winded, but perhaps the little-known historical aspect will be of some interest.

First, the primary navigation and guidance computer for the Apollo missions was called the AGC = Apollo Guidance Computer and, as the questioner states, was located physically in the Command Module (CM) and in the Lunar Excursion Module (LEM). There were many other computers elsewhere (as in the Saturn booster) intended for secondary backup navigation and guidance in the event of failure of the AGC's.

As far as where to get detailed information on the AGC's, the question is a little ambiguous. Does the questioner want detailed information (e.g., program listings) on the programs themselves, or does he want to know more about the architecture (which was VERY unusual) of the AGC's?

Unfortunately, I cannot help much with the former. Though I wish now that I had, I did not save any listings of the onboard programs. However, I can provide a little information about the architecture.

And, here is where a very interesting bit of little-known history comes into play.

It turns out that the Apollo primary guidance and navigation computer (AGC) was actually derived DIRECTLY from a much earlier (than the Apollo project) proposal by the Instrumentation Laboratory (now Draper Lab) at MIT entitled "A Recoverable Interplanetary Space Probe, Report R-235" I happen to have a copy of that proposal, which spans 4 volumes as follows:

Vol. I - General Discussion (pp. 1-66)

ch. 1 Introduction

ch. 2 A Description of the Recoverable Interplanetary Space Probe Experiment

ch. 3 A Plan for an Experimental Program

Vol. II - Engineering Analyses (pp. 67-312)

ch. 4 Fundamentals of Round Trip Interplanetary Trajectories

ch. 5 Theory of Space Ship Navigation

ch. 6 Functional Configuration and Operation of Space Vehicle

ch. 7 Structural - Thermal Configuration

Vol. III - Engineering Analyses (pp. 313-570)

ch. 8 Electrical System

ch. 9 Description of Accessories

ch. 10 Computer

ch. 11 Micro Rocket

ch. 12 Communication System

ch. 13 Re-entry Vehicle Design

ch. 14 Photographic Prospect

ch. 15 Radiation Environment

Vol. IV - Appendices (pp. 571-810)

A The Determination of Round-Trip Planetary Reconnaissance Trajectories

B General Theory of the Navigational Fix

C The Fundamental Perturbation Matrices

D Formulation of a Navigation Theory

E Midcourse Navigation Error Analysis

F Error Analysis for Planetary Approach

G Procedure for Computation for the Fundamental Matrices

H Space Vehicle Computer Programming

I Disk Scanning (Appendix I refers to scanning the disk of a planet for navigation purposes, and not to scanning a computer disk!)

I have taken the trouble to list the individual chapter names because they reveal that MIT had, by the time of its proposal, worked out all of the major aspects of navigation between celestial bodies within the solar system. The intention was to recover high resolution reconnaissance photographs of a nearby planet, such as Mars or Venus (ch. 1). However, the remarkable thing is that the date of this proposal was July, 1959!!!!!

Not only had MIT worked out the details of a navigation scheme, but the computer architecture suggested in the 1959 proposal was subsequently adopted nearly wholesale for the AGC. This is, as far as I can tell, a little known fact and it illustrate how early a given technology gets locked in when a very complex task is involved.

And, one of the most unusual aspects of the AGC, taken directly from the 1959 proposal, was the use of a so-called "rope memory" for fixed data/program storage. In this memory, each computer "word" is stored by a single magnetic core. If the word is to consist of 16 bits, then 16 wires, each corresponding to a bit of the word, are passed either through or around a magnetic core. When the core magnetization is flipped, a pulse (a "1") appears on the wires that passed through the core, and no pulse (a "0") on the wires that passed around the core. A second computer word is created by threading the same 16 wires in a different pattern, depending on the desired content of the word, either through or around a second core. To read out the second word, the magnetization of the second core is flipped, and the pattern of pulses read from the same set of 16 wires. When the entire set of words/cores is in place, the whole thing resembles a "rope" with a string of cores all along its length, each corresponding to a unique word of fixed memory.

This very memory architecture was used in the AGC, primarily because it was considered absolutely recoverable in the event of, for example, interference from severe cosmic rays. No cosmic ray would ever be capable of altering the fact of a given wire passing through or around a give core!!!

I can also tell you that the AGC processor used RPN (Reverse Polish Notation) in which operands are first pushed onto a stack and then popped off as they are operated upon by various operators. This is, of course, an architecture that is seen in many modern processors, such as the floating point processor used in PC's.

I would suggest that this very early 1959 MIT report would be an excellent place to start to learn about the AGC computer. I'm sure that the report has been preserved in the MIT archives [Institute Archives and Special Collections: 617-253-5136]. It is also likely that the MIT Archives contains other similar follow on reports more directly related to the AGC.

Finally, for what it is worth, recently (a few years ago) I discovered quite by accident that the Computer Museum in Boston has (or at least then had) a genuine AGC computer on display. Perhaps other computer museums around the country may have obtained AGC's for display, enabling the questioner at least to get a look at one.

I hope that the above is at least helpful, if not definitive.

In addition, I had a poke around the historical sections of the the NASA HQ site. There is a report on the history of computers in spacecraft there including a chapter on Apollo, focusing especially the AGC. By about the fifth page it starts to get pretty detailed, and the list of references is quite extensive. I'll bet many of them are in the MIT Institute Archives and Special Collections that Warren points to.

Hope this helps. If you need more, please ask and I'll see what I can come up with.

David Ehnebuske